Chapter 1

UNLIKELY REBELS

Itinerant Experts

Clothed in a light blue T-shirt and chinos, Kent Cox revealed nothing tosuggest that he was in the vanguard of an employment revolt that wasspreading quietly across America’s industrial landscape.1 Kent had noobvious tattoos or piercings. He carried no union card. He was not outto change the world. His passions were for code, ballroom dancing, andscience fiction, not for politics—especially not for politics of the organi-zational sort. His short, dusty blond hair was only slightly rumpled, andhis wire-rims suggested nothing more than intelligence. Had you seenhim that day sitting across from us in a Chinese restaurant in downtownMountain View in the heart of Silicon Valley, you might not have givenhim a second glance, so well did he blend in with the other techies whofrequent the establishment in search of noonday egg rolls and sweet-and-sour soup. If you had noticed a difference between Kent and the rest, itwould have been in the minutiae: he entered the restaurant alone, not asa member of a pack, and his T-shirt bore no company logo.

On the other hand, if you were a programmer who had been in the Sil-icon Valley long enough, you might have recognized Kent, at least byname, as a guru of sorts. Firms on the cutting edge of high technology ea-gerly sought Kent’s expertise as a software developer. At thirty-six, hehad already worked for a “who’s who” of companies, some of which arehousehold names across America; others are famous only among thetechnogensia. Kent’s six-figure income was testimony to his talent. Yet,by some people’s measure, he had not been gainfully employed in nearlya decade, if being gainfully employed means being permanently em-ployed. Since 1989 Kent had worked only as an independent contractor,a “hired gun” as they are sometimes called, moving from firm to firmevery six to eight months in search of another challenge.

1 Because a number of people appear repeatedly over the course of our narrative, weprovide a “Cast of Characters” in the appendix, to which readers may turn if they need areminder of the individual’s affiliation, role, occupation, and involvement with contracting.In this way we avoid the awkwardness of introducing our informants more than once in thecourse of our telling.

At sixteen, when most of his peers were still entranced by the dancingphosphorescence of Space Invaders, Kent began programming. He en-rolled in Cal Poly a year later, where he majored in computer science.Hearing no scholarly calling, Kent took several years off to write code. “Iwas an indifferent student,” he admitted. “I was only motivated to pro-gram.” After eventually graduating in the mid-1980s, he programmedfor his father’s start-up for several years, and in 1989 he went to work asa full-time developer for Sun Microsystems. After a year of eighty-hourweeks, Kent “burned out,” left Sun, and took an eight-month vacationto, as he put it, “put my head back together.” He realized he didn’t wantanother full-time job. Indeed, he was ambivalent about work in general.But in 1990 a friend convinced him to join a software project as a con-tractor. From then on Kent was never permanently employed, yet he hadworked continuously on a stream of projects that ranged from buildingdevice drivers for printers and digital cameras to developing applicationsfor the Web. He attributed his success not only to his skills but also to hisreputation in the technical community. “My career,” he explained, “hasbeen not so much getting hired to do the same thing, but getting hired bypeople I know who can’t find someone they need to fill a particular niche,so they hire someone they think can learn to do it.”

Like many experienced contractors, Kent found work through col-leagues, friends, and acquaintances. “When you want to get work,” heexplained, “the key is to make sure everybody you know knows you’relooking. Just call all of your friends and say, ‘Hey, I’m looking and one of‘em will say, ‘Oh we need somebody to do IO [input-output].’ NEC camethrough somebody I danced with.2 Her husband was doing contractingfor them. You get contracts anywhere. Anyone you know is a potentialcontract. One of my friends gets a lot of his jobs through church.” Kentwas currently charging $90 an hour for his services and working thirtyhours a week. He could have worked more, but he preferred to use therest of his time to surf, read, and write his own software.

At the time we met Kent, he had just run up against one of the catch-22s of being a hired gun. He had contracted to provide guidance to aproject team supposedly composed of “crack engineers.” But, he ex-plained, “I have not meshed well with this team. Most of them are ontheir first or second project. I’m more like on my twenty-fifth, and thatgives you a very different perspective. I did some egregious political blun-dering in the beginning when I pointed out things that were going to beproblems in three months. They did become problems, but I was askingfor trouble because they couldn’t see what was coming. I’ve gotten backinto everybody’s good graces, but it’s been a strain. This is the first time

2 CHAPTER ONE

2 NEC was a company for whom Kent had contracted.

I’ve really had this problem because I’d been working for a long timewith people who have a similar level of experience as me.”

Kent’s only regret about having worked so long as a contractor wasthe proverbial Silicon Valley story of opportunity lost: “One of the hard-est things is watching people regularly get rich. I turned down a job atNetscape because I wanted to contract.3 I mean, I turned down a joboffer from Netscape the month before they went public! I decided not toapply when they moved to Mountain View even though I was doing Webdevelopment at the time. That’s the one that hurts. Of course, I had noidea at the time that Netscape was going to do that. I keep remindingmyself of that when I start feeling really stupid.”

Earlier in the year, Kent had founded the first Java users’ group in theValley, which he described as a platform for advancing as a contractor.“I’ve decided that I should probably think about moving to the nextlevel,” he confided, “where people are hiring you as a name rather thanjust hiring you as a warm body or a coder. Stenogram has the guy thatwrote the book on Microsoft mail’s API [application program interface]on retainer. We tried to get the guy that wrote the book on programmingfor Microsoft on retainer as well. That sort of public recognition was oneof the reasons I started the users’ group and that also means that peoplecontact me. Java people send me mail. Microsoft invites me up to Red-mond to look at things. I just figured it was a way of putting myself inthe way of job opportunities.” In fact, we were introduced to Kent pre-cisely because our contacts told us that he was central in the emergingnetwork of Java programmers.

If Kent Cox was an unlikely foot soldier in an employment revolt,Yolanda Turner, a well-dressed, fifty-two-year-old African-Americanwoman with three grown children and an infectious laugh, was even lesslikely. Yolanda had lived her entire life in and around Los Angeles. In1982, at the age of thirty-six, Yolanda decided to pursue a bachelor’s de-gree in film at Pomona College after raising her children as a singlemother. Four years later, she graduated and worked a number of tempo-rary clerical jobs before landing a full-time job, first at UCLA and then atRockdyne, as an administrative assistant. Had Yolanda’s story endedhere, she might have served as a classic example of how difficult it is forolder African-American women to rise above socioeconomic and racialbarriers, even with a degree from a prestigious college. But Yolanda’s su-pervisor at Rockdyne noticed her interest in PCs and paid for her to takecourses in hardware and software. Before long, she became an in-houseexpert who informally assisted and coached her coworkers when their

UNLIKELY REBELS 3

3 From today’s perspective, Kent’s decision not to go to work for Netscape does not ap-pear as bad as it did at the time.

computers went out of whack. Gradually, Yolanda began to spend mostof her time fixing workstations and tweaking software, even though shewas still employed as a secretary. Six years later in 1995, and still em-ployed as an administrative assistant, Yolanda submitted her résumé toan online job board and landed a contract as a workstation technician atan IBM subsidiary. Since then Yolanda had worked exclusively as a con-tractor, moving over time from workstation maintenance to systems ad-ministration and then on to quality assurance and testing.

Unlike Kent Cox, who relied on his extensive network, Yolanda se-cured her positions solely through staffing agencies that served as bro-kers between her and the clients who bought her services. These agencieshad sprung up in the 1980s and 1990s to take advantage of firms’ lack ofinformation about contractors and contractors’ lack of informationabout job openings. Most, but not all, modeled themselves after tempo-rary help agencies. In return for locating positions, deducting payrolltaxes, and occasionally offering access to health insurance, agenciesadded a markup of approximately 30 percent to 40 percent to Yolanda’shourly pay of $35 per hour. Yolanda didn’t much like the markup, whichshe considered excessive, but she accepted it as the way the game wasplayed. When we met Yolanda in 1998, she was working forty- to fifty-hour weeks for an annual income of about $87,000. Being a morningperson, Yolanda typically arrived at work at 6:00 A.M. and left at 2:30 inthe afternoon. Often, she worked part of Saturday or Sunday as well.Yolanda’s contracts typically lasted between three and six months, andsince 1995 her total “downtime”—the time in between contracts—amounted to less than two weeks. When asked why she had forsakenpermanent employment, Yolanda left no doubt about her motives: “Sig-nificantly more pay! Significantly more independence! And, you don’t getimmersed in the political environment that goes on in companies. I likethe independence. I like being able to move around and see how differentcompanies do things. In most cases you make a significantly highersalary than you do as a full-timer.” Yolanda estimated that she typicallymade $10 to $15 more an hour than permanent employees doing thesame work. When she first started contracting, however, this was not thecase: “I used to do work on the side fixing people’s computers and I wascharging $15 an hour and I thought, ‘I’m doing pretty good.’ Then, I meta girl who was charging $65 an hour and I said, ‘Wait a minute. What’swrong with this picture?’ So, I started charging more and getting it. Tomake it as a contractor,” she advised, “you need to know what the mar-ket is and what your skills are worth.” Yolanda did this, in part, by talk-ing with other contractors and, in part, by subscribing to Contract Pro-fessional, a monthly magazine targeted at contractors.

Yolanda had mastered the art of staying current. At the time we spoke

4 CHAPTER ONE

with her, she was learning SAP with the goal of becoming an SAP con-sultant within three years.4 She had chosen SAP because “it’s very lucra-tive, and that’s what I’m focusing on: money.” In addition to buying andreading technical books and surfing the Web for more up-to-date infor-mation, Yolanda maintained her own home computer lab. “I have a fullsystem at home. I’ve got two computer systems and two different typesof modems. I’ve got a scanner and two printers and a little digital cam-era. So, I have all the equipment I need to learn something, if I need toknow it.” Yolanda chose her contracts for their learning potential. “Idon’t have a lot of time to read,” she reported, “so while I’m working onthis product—and it’s a product that’s in SAP—I’m learning as I goalong. I’m pretty familiar with about three of the modules right now.And in some cases,” she added, laughing, “more familiar than some ofthe people who are the SAP consultants there.”

At first glance Julian Stoke could not have been more different fromYolanda Turner. Julian was a corpulent white man with thinning hairand horn-rims. To call him disheveled would have been an understate-ment. He wore wrinkled chinos, running shoes, and a short-sleeved plaidshirt whose tails had sneaked out on both sides of his pants, revealingtwin triangles of pale white skin. A devout Mormon, Julian was the fa-ther of seven children ranging from seven to twenty-four years of age. Al-though Julian was part of the same movement in which Kent andYolanda had enlisted, his story was different.

After graduating from Brigham Young University in 1974 with a self-designed bachelor’s degree in computational linguistics, Julian movedwith his wife and first child to the Silicon Valley to be near his parents.Within a year he was working full-time for The Gap, where he climbedthe ladder over the next five years from programmer to third-level tech-nical manager. He was working fifty- to sixty-hour weeks and like KentCox finally reached the point where he burned out. But rather than taketime off, Julian wended his way through three more permanent positionsin rapid succession until he finally landed at Citicorp in 1981 as a sys-tems programmer. Before his probationary period ended, however, hewas fired.

“They quit me!” Julian explained as if still reeling from incredulitytwenty years later. “The irony was that the manager of the group was acontractor. He came to me one day and said, ‘You don’t fit in with our

UNLIKELY REBELS 5

4 SAP is the most widely used inter-enterprise software program developed and mar-keted by the German firm of the same name. Inter-enterprise software is a class of programsthat integrate data across a variety of application areas including finance, human resources,customer relationships, supply chain management, forecasting, and so on. It is a kind ofone-stop software platform for firms.

group.’ I said, ‘Right, I agree with you. You are not the same kind of peo-ple that I am.’ He said, ‘We will have to let you go.’ I said, ‘Well, have Idone something wrong? You don’t like my work?’ and he said, ‘No, youjust don’t fit in.’”

After looking for work unsuccessfully for “quite a while,” Julian casthis lot in with a friend, an independent contractor with whom he hadworked at The Gap. Like Yolanda and Kent, Julian found the lure ofmoney and autonomy hard to resist. “I am not ordinarily a ‘live-on-the-edge’ sort of guy, but I was really intrigued by several aspects of con-tracting: the money, the mobility, and the professionalism.” For most ofthe 1980s, Julian sold his services as a VM system programmer to anarray of clients in the Bay Area through his friend’s business.5 But in1988 the business collapsed and his friend absconded with three thou-sand of Julian’s dollars. Julian then took his first contract through a bro-ker and had since worked exclusively through staffing agencies. “Onceburned, twice shy,” he explained.

Over the years, Julian had become savvy about dealing with agencies.“Every broker tells you, ‘You have a career with us, and when we aredone with this assignment, we’ll will find you another. You’re a valuedmember of our team.’ But what it really comes down to is that if theyhave no openings, there is no commitment to you.” To cope with uncer-tainty and the lack of commitment, and to increase his chances of findingwork when he needed it, Julian had learned to work with many agenciessimultaneously. Although he did not like it, he accepted their markup—usually around 40 percent, he told us—as an unavoidable fact of the con-tracting life. Several years earlier, Julian had gone through a trying pe-riod without health insurance. At one point he was even forced to fallback on the charity of the Mormon Church to help cover steep medicalbills incurred after his son was injured in a road accident. Julian subse-quently took a permanent job for a year, largely to qualify for the HMO,which he was able to retain at his own expense once he returned to con-tracting.

Unlike Kent and Yolanda, Julian’s career as a “warm body” had beenrocky. Since 1988, he had moved between contracting and permanentemployment on four occasions. Although he had seen little downtime, hehad been fired on three occasions, once from two separate contracts si-multaneously. When we met Julian, he was making $55 an hour, work-ing over sixty hours a week (not including the time he spent keeping hisskills up-to-date), and having difficulty making ends meet. He blamed hisfinancial difficulties on old debts, on a wife who was unable to work, andon being “bad with money,” especially credit cards. Julian had not taken

6 CHAPTER ONE

5 VM (virtual machine) is an operating system for IBM mainframes.

a vacation in years. With such wages, “It’s a huge temptation to workevery hour of the day,” he admitted.

Julian, like Kent, was involved in a professional community. Butwhereas Kent’s was formally organized, Julian’s existed loosely on theWeb. “Do you know about the IRC?” Julian asked in response to ourquestion about professional relationships. “That’s one of the major typesof chat rooms, where you get online to talk and make friends. I have onegroup of professional people I get on with every once in a while. That isone of the tools of problem solving. When you can’t find the answer to atechnical problem in the manuals, the next best thing is use a bulletinboard. I subscribe to about ten of them. I go on and put out my question,‘Why is this happening? How do I get around this?’ I get answers all thetime. Chat has changed the face of contracting in recent years, becausewhen people in a company have problems they ask the contractor, whomthey see as a guru, their mentor.”

Although Julian’s wife had misgivings about contracting and periodi-cally urged him to “get a real job,” Julian was reluctant to do so. His lastfew contracts had gone well, and he had doubts about his ability to holddown a permanent job. The opportunities to learn new technologieswere plentiful. And, like Kent, Julian saw himself as an expert, a systemsprogrammer who saved his clients from themselves. “We contractorshave a superiority, a superciliousness,” he confided. “We sometimes getthe impression that we are doing the real work, that we are going the dis-tance. These other guys go home at five and we stay until eight. But wealways know that there will come a day when we sign our project over tosomebody and it is no longer our problem. It’s theirs.”

Despite their many differences, Julian, Yolanda, and Kent had onething in common: all had made a conscious decision to step outside themainstream of employment. They had questioned and then jettisoned, atleast for the time being, the vision to which their parents and society hadtaught them to aspire: the safe haven of a permanent job. Like most ofus, they had been told that with a college education they could get a“good job,” a code word for going to work full-time for a reputable em-ployer. In the absence of a major recession and in return for loyalty andhard work, employers would reward them with a good salary, health in-surance, promotions, and perhaps most important, reasonable expecta-tions for long-term security. In the end, they’d retire at sixty-five with apension, a gold watch, and some savings.

At least initially, like most Americans their age, all three had reason tobelieve in the bargain. They were, after all, technical professionals whomemployers historically valued more highly than most other kinds of em-ployees. But for reasons ranging from being laid off to longing for a lesshectic lifestyle, each had decided to bet on his or her skills and forgo full-

UNLIKELY REBELS 7

time work, even though all believed they could have secured a permanentjob had they so desired. Each knowingly entered a world of work thatdemanded they change employers two to three times a year and adopt anotion of security based on their continued ability to market their ex-pertise to a portfolio of buyers. In short, they had left behind the famil-iarity of traditional employment for the uncertainty of a more fluid labormarket. None seemed to regret their choice.

Kent, Yolanda, and Julian are what we call technical contractors. Theysometimes even called themselves contractors (or consultants), althoughthey usually introduced themselves by their occupation: software devel-oper, technical writer, multimedia developer, systems programmer, andso on. As contractors, they possessed skills and knowledge similar tothose of the full-time engineers, software developers, and technical writ-ers with whom they worked. They performed similar tasks and workedin the same firms as the full-timers. Yet, even though they worked at thecutting edge of the postindustrial revolution, the mode of their employ-ment resembled that of craftsmen in the Middle Ages, especially thestonemasons who built the great cathedrals of Europe.6

Stonemasonry was, and still is, the queen of the crafts. Although wedon’t think of stonemasons as high-tech workers, in medieval Europethey were. Shaping artful and well-engineered Gothic cathedrals fromhand-hewn stone was a significant technical achievement requiring ex-pertise that took years to master. For this reason, stonemasons were inshort supply and great demand. Building a cathedral was an expensive,multiyear project commissioned by the clergy and royalty of a city. Sincemost medieval cities were small, they needed and could only afford onecathedral. Accordingly, masons could expect to be gainfully employed inone place for no more than a few years. Because of this constraint,stonemasons worked as itinerant craftsmen, traveling from city to citythroughout Europe, joining projects and leaving cathedrals in theirwake. Like these artisans of old, technical contractors also ply their tradeas “itinerant experts.” Some, known in contracting circles as “nomads”or “road warriors,” even move from city to city in search of projects, justas stonemasons did five centuries ago. Firms substitute for cities andprojects for cathedrals.

In late 1997 when we set out to study technical contracting, we didn’tthink of ourselves as entering a world of itinerant experts. We certainlydidn’t think of contractors as social pioneers. Our goal was simply to un-derstand how employment relations were changing at the dawn of the

8 CHAPTER ONE

6 Knoop and Jones (1967), Gimpel (1980), and Applebaum (1992) describe the work ofmedieval stonemasons and the organization of cathedral building in medieval Europe.

twenty-first century. We hoped to document the social dynamics of skilled“contingent labor,” a term economists and sociologists now use for anarray of short-term arrangements including part-time work, temporaryemployment, self-employment, contracting, outsourcing, and home-basedwork. To be honest, if anything, we saw contingent work, in general, andpossibly even contracting, as social problems indicative of corporateAmerica’s willingness to abdicate its responsibility to employees in returnfor greater profits. Contingent employment seemed to bear a troubling re-semblance to the laissez-faire capitalism of the late nineteenth century,which, in the name of individualism and the free market, spawned aplethora of exploitative practices.7 In our concern we were hardly alone.

The Unraveling of Permanent Employment

The foundations of the American system of employment were laid afterthe Great Depression. The system was conceptualized under the NewDeal and then institutionalized in labor law and collective bargaining.Permanent employment was the system’s cornerstone. It hinged, in largemeasure, on a bargain struck between employers and employees. Ac-cording to this bargain, as generations of workers grew to understand it,employers would fulfill legally and culturally prescribed obligations toemployees. These included offering reasonable expectations for job secu-rity and benefits, such as health insurance and pension plans. In return,employees were expected to offer their loyalty and their best efforts.

By the end of the 1990s, many Americans feared that this bargain andthe job security it brought were unraveling before their eyes.8 A number ofinterrelated developments contributed to the apprehension. The first, andmost troubling, was “downsizing,” a bit of forked-tongue corporate-speakthat came to replace the older and less Orwellian vocabulary of “layoffs.”Until the mid-1980s, everyone thought they understood layoffs. Lay-offs were the bad luck that happened to blue-collar and clerical workers

UNLIKELY REBELS 9

7 Laissez-faire capitalism refers to the doctrine that economic systems function bestwhen they are free of governmental interference. Having first been articulated by Frenchscholars, John Stuart Mill, Adam Smith, and Jeremy Bentham took the doctrine of laissez-faire as the central tenet of classical economics and fused it with a doctrine of radical indi-vidualism. In the early twentieth century, laissez-faire practices were associated with therise of monopolies and cycles of boom and bust. Laissez-faire principles still underwriteconservative economic doctrine in the United States and Great Britain. For a superb discus-sion of laissez-faire capitalism in the United States and Great Britain, see Bendix (1956).

8 Peter Cappelli, Paul Osterman, and Thomas Kochan have written extensively aboutthe unraveling of the New Deal and its institutions (Kochan, Katz, and McKersie 1986;Cappelli et al. 1997; Cappelli 1999; Osterman 1996, 1999; Osterman et al. 2001).

whenever the economy headed south or a firm experienced financial troublefor other reasons. But no matter how bad things got for men and womenwith blue, pink, or no collars, most white-collared professionals and man-agers could assume that their jobs were relatively safe, even in hard times. Itwas partially for this reason that the Ozzie and Harriet generation ofAmerican parents exhorted their children to get a “good education.”

During the 1980s, the rules of the game seemed to change. For thefirst time in history, firms began to shed professional, technical, andmanagerial workers in large numbers. In fact, by the mid-1990s corpo-rate downsizings were more likely to target managers and professionalsthan they were blue- and lower, white-collar workers.9 Even more per-plexing to Middle America was that downsizings, unlike layoffs of thepast, seemed independent of economic cycles. The top management ofsome firms publicly admitted to resorting to downsizing as a way ofboosting their firms’ stock price.10 By the late 1990s, the consequencesof the new rules of employment finally showed up in aggregate laborstatistics: job tenure among white males had begun to fall for the firsttime since World War II.11 Furthermore, despite a vibrant economy,people’s levels of trust in their employers sank as feelings of insecurityrose.12 Even managers and professionals who had not yet been hit bydownsizing were apprehensive.13 The business press and career coun-

10 CHAPTER ONE

9 Moore (1996), the American Management Association (1996), the New York Times(1996), and Farber (1997) provide discussions of and data on downsizing in the mid-1990sand its effects on workers. Academic research on downsizing has mushroomed over the lastdecade. Rousseau’s work (Rousseau 1995; Rousseau and Anton 1991) is particularly en-lightening with respect to employees’ perceptions and expectations.

10 The most vocal and notorious champion of downsizing was Jack Welsh, the CEO ofGeneral Electric. He was known among employees at GE as “Neutron Jack”: like the neu-tron bomb, Welch eliminated people and left the buildings intact.

11 Until recently most economists found little evidence of a general decline in job securitywhen measured as average job tenure (Diebold, Neumark, and Polsky 1997), but this maysimply have reflected the fact that analyzable data lagged reality. In 1998, new data com-piled by the Bureau of Labor Statistics (BLS) indicated for the first time that job tenure haddeclined among American men for all but the youngest and oldest age cohorts (Bureau ofLabor Statistics 1998).

12 For data on increasing feelings of insecurity and distrust of management during the1990s despite stable levels of job satisfaction, see table 2.3 of the National Research Coun-cil’s (1999) The Changing Nature of Work.

13 Katherine Newman (1989) and Charles Heckscher (1995) insightfully document thestigma and fear felt by managerial and professional workers who were downsized in the1980s and 1990s. Comparing the lot of blue-collar workers who had been laid off by theSinger company with the experiences of downsized managers led Newman to the conclusionthat white-collar workers are more likely to experience depression, alcohol abuse and otherpersonal problems. She argues that managers fare worse because their friends and family seebeing laid off as a personal failure. The friends and family of blue-collar workers understandthat being downsized is simply something that firms do to people for economic gain.

selors began advising managers to treat security as an outdated pipedream and to focus, instead, on “employability”—the attitudes, behav-iors, experiences, and skills that would help them land the next job.14

The message was clear: in the future, a lot fewer people were going tobe earning gold watches. Loyalty was passé and self-reliance was invogue.

As if downsizing was not enough of a change, during the 1980s employers increasingly began moving jobs offshore to countries wherelabor was cheaper. In the name of economic agility, firms also began to outsource goods and services that they formerly provided for themselves. The logic of outsourcing dictated that it was often cheaperto buy goods and services from suppliers than it was to provide them internally. Finally, and for our purposes most important, employ-ers began to make greater use of contingent workers—individualshired, often through staffing agencies, for a limited period of time to perform a specific job. Although firms had long employed tempo-rary workers for seasonal and short-term needs (for example, to staffretail stores at Christmas or to stand in for absent full-timers), duringthe late 1980s corporations began to view temporary labor as an ex-tension of the broader strategy of outsourcing. Outsourcing labormade it easier for firms to shift their employment mix because con-tracted services were easier to terminate on short notice. Moreover, hiring contingent workers relieved managers of some of their more dif-ficult tasks, such as worrying about thorny “people issues” or cateringto the motivations and feelings of employees. Business strategists, suchas Charles Handy, argued that flexibility was key to survival in a globaleconomy. One way to enhance flexibility was to surround the firm’score employees with a buffer of contingent workers who could be hiredand fired at will.15 Business leaders were apparently convinced. Howthey went about accomplishing this vision, depended, to a great extent,on the laws that governed and regulated employment. The fact thatYolanda and Julian, like most technical contractors, found their jobsthrough staffing agencies was one upshot of the legal wrangling thatensued.

UNLIKELY REBELS 11

14 For a sense of the rhetoric of employability, see Bridges (1994), Carlin (1997), andCaulkin (1997).

15 Handy’s (1989) book, The Age of Unreason, was a best-seller. The academic literatureindicates that firms frequently hire contractors for the reasons Handy promoted. For em-pirical evidence on firms’ motives for hiring contractors, see Mangum, Mayall, and Nelson(1985), Abraham (1988), Pfeffer and Baron (1988), Hakim (1990), Davis-Blake and Uzzi(1993), Abraham and Taylor (1996), Matusik and Hill (1998), and Kalleberg andReynolds (1998).

The Legal Context of Contingent Work

American tax and employment law has long rested on two assumptions.The first is that people are either self-employed and, hence, work forthemselves, or else they are employed by the organization for which theyperform tasks. The second assumption is that people traditionally haveone employer at a time, and even when they have more than one em-ployer, as in the case of moonlighters, their relationship with each re-volves around a separate and distinct task.16 Lawmakers have drawn onthese notions to assign tax liabilities and responsibilities, to distributeand enforce workplace rights, and to secure the American system of so-cial welfare.

U.S. law requires that employers pay employment taxes on every em-ployee and that they also withhold the workers’ employment taxes fromthe workers’ paychecks. In short, employers serve as the government’stax collectors. Employment taxes include federal, state, and local incometaxes, social security contributions, and federal unemployment tax.Under the Fair Labor Standards Act, employers must pay their workersat least the minimum wage and meet overtime obligations. Title VII ofthe Civil Rights Act forbids employers from discriminating against em-ployees on the basis of race, gender, or creed. The Age Discrimination inEmployment Act adds age to the list of forbidden discriminations, whilethe Americans with Disabilities Act extends protection to otherwisequalified employees with physical disabilities. The Family and MedicalLeave Act guarantees people that they will have time for handling familyand medical emergencies by requiring employers to grant employees upto twelve weeks of unpaid leave per year when such events occur. TheEmployment Retirement Security Act sets parameters for employee pen-sion plans, including how much service is needed before an employee isvested. Finally, the National Labor Relations Act grants employees theright to organize and governs aspects of labor-management relations.

In short, employers have significant obligations to and for their em-ployees under U.S. law. But employers incur none of these obligationswhen they use the services of an independent contractor. Independentcontractors are considered self-employed persons, even if they are not in-corporated as a business. Therefore, employers who wanted to outsourcelabor originally had significant incentives to make use of independent

12 CHAPTER ONE

16 Our discussion in this section draws heavily on Muhl’s (2002) useful paper, “What Isan Employee?” Other useful sources on contracting’s legal environment include Jenero andSpognardi (1995), Everett, Spindle, and Turman (1995), Fox (1997), Carnevale, Jennings,and Eisenmann (1998), and Zeigler (2001).

contractors, because doing so absolved them from a series of costs andobligations. From the perspective of the federal government, however,any shift from traditional employment to the use of independent con-tractors represented a potential loss of taxes. It also undermined thelaw’s primary mechanism for redressing discrimination and ensuring thewelfare of the labor force.17 For this reason, the law allows the govern-ment, especially the Internal Revenue Service (IRS), to prosecute andseek damages from employers who try to avoid their legal duties. Thelaw also allows workers to seek compensation and additional damageswhen they believe that a firm has classed them as an independent con-tractor to avoid paying appropriate wages and benefits.

During the 1980s, employing technical contractors became popular asfirms discovered that they could reduce labor costs by substituting inde-pendent contractors for employees. In some cases, firms actually dis-missed former employees and then rehired them as independent contrac-tors. In other instances, firms hired people who were clearly independentcontractors, but then treated them as permanent employees. For in-stance, firms might deny contractors’ rights to the intellectual propertythey created (which is illegal) or demand that they work exclusively forthe firm as a condition of employment. Microsoft became legendary foremploying “permatemps,” although substituting independent contrac-tors for employees was widespread among high-technology firms.

Prior to 1990, Microsoft routinely hired a large number of indepen-dent contractors, many of whom were former employees. Microsoft re-quired the freelancers to sign agreements stating that they were indepen-dent contractors, not employees, and that nothing about the agreementcould be construed as implying an employee-employer relationship. De-spite the agreement, Microsoft made no distinction between contractorsand full-time employees, except that the former were hired for specificprojects. Contractors worked the same hours as full-timers at the samelocations and were supervised by the same managers in the same way.Yet, Microsoft did not pay the contractors’ employment taxes and re-fused to allow the contractors to participate in the company’s pensionand stock plans.

Concerned that firms like Microsoft were using independent contrac-tors as a way to avoid taxes, in 1987 the IRS announced a new test bywhich it would decide whether a worker qualified as an independentcontractor.18 The test, based on a common-law definition of employment,

UNLIKELY REBELS 13

17 Because independent contractors are required to pay both the employer’s and the em-ployee’s portion of all employment taxes, replacing an employee with an independent con-tractor should not reduce the government’s total tax income. However, it is more difficultto monitor the compliance of independent contractors than it is to monitor employers.

18 Revenue Ruling [87-41, 1987-1 CB 296].

rapidly became known in contracting circles as the IRS’s “twenty-ques-tion rule.”19 Table 1.1 displays the twenty questions and the responsesthat the IRS expected of an independent contractor.20 Although the IRSheld that the answer to no single question was sufficient to grant or denythe status of independent contractor, long-term contractors like JulianStoke agreed that after 1987, it became increasingly risky to proclaimoneself an independent without being incorporated.

In 1989, the IRS used the twenty-question rule to go after Microsoft. Ina highly publicized ruling, the IRS found that Microsoft had illegally mis-classified workers as independent contractors to avoid payroll taxes. Mi-crosoft accepted the IRS’s ruling, paid a stiff penalty, and subsequentlyconverted some of its contractors to employees while dismissing manyothers. Feeling cheated, those whom Microsoft dismissed filed a class ac-tion suit (Vizcaino v. Microsoft) claiming that they were entitled to partic-ipate in Microsoft’s pension and stock plans because Microsoft hadtreated them as employees, not as contractors. The case moved quicklyfrom state to federal court, where Microsoft initially won the case, but theplaintiffs appealed. In 1995, the U.S. Court of Appeals for the Ninth Cir-cuit reversed the lower court’s decision, finding in favor of the plaintiffs.The court held that because Microsoft had treated the contractors as em-ployees, they were entitled to pension funds and stock options. Microsoftmanaged to delay final settlement until 1999, when the Ninth CircuitCourt brought the case to a close by ordering Microsoft to distributenearly $9.7 million among eight thousand to twelve thousand people.21

The IRS’s decision to prosecute Microsoft and the Ninth CircuitCourt’s ruling sent a shock wave throughout the high-technology indus-try. Firms suddenly became wary of hiring independent contractors. Toavoid the possibility of finding themselves in conflict with the IRS or inviolation of employment law, lawyers advised firms to insist that even in-corporated contractors work through staffing agencies, which wouldserve as the contractors’ “employer-of-record.” The changes in howfirms approached contracting benefited the staffing industry, which grewin leaps and bounds along with the demand for contractors.

14 CHAPTER ONE

19 A common-law definition of employee pivots on the notion of agency, “which, in anemployment context, consists of a relationship wherein one person (the employee) acts foror represents another (the employer) by the employer’s authority” (Muhl 2002, 5).

20 It is important to understand that circumstances dictate whether any question is appli-cable. The IRS treats no question as sufficient for identifying an independent contractor.Nor does the IRS demand that an individual be able to answer all questions in the expectedway before granting the status of an independent contractor. Instead, the IRS uses thetwenty questions as a general guideline for classifying the facts of a case.

21 The text of the Ninth Circuit Court’s final ruling can be found at www.techlawjournal.com/courts/vizcaino/19990512.htm.

UNLIKELY REBELS 15

Table 1.1The IRS’s “Twenty Questions” for Distinguishing between Employees and

Independent Contractors

Response Indicating anQuestion Independent Contractor

Must the worker follow the company’s instructions about when, where, and how to work? No

Does the company provide the worker with training? NoAre the worker’s services integral to the business? NoCan the worker subcontract the work to

someone else? YesDoes the company hire, supervise, and pay the

worker’s assistants? NoIs the worker employed for an extended, continuous

period? NoDoes the company set the worker’s hours? NoMust the worker work full-time for the company? NoDoes the worker perform his or her services on the

company’s premises? NoMust the worker perform services in an order or

sequence set by the company? NoMust the worker submit oral or written reports to

the company as part of his or her evaluation? NoIs the worker paid by the hour, week, or month? NoDoes the company pay the worker’s business or

traveling expenses? NoDoes the company furnish significant tools, materi-

als, and equipment? NoDoes the company have the right to fire the worker? NoCan the worker quit without incurring a liability to

the company? NoDoes the worker have a significant investment in

tools or facilities? YesCan the worker realize a profit or loss as a result of

his or her services? YesDoes the worker provide substantial services to

multiple companies at one time? YesDoes the worker regularly advertise or make his or

her services available to the general public? Yes

Estimating the Size of the Contingent Workforce

By 1997, when we began our study, contractors like Kent, Yolanda, andJulian made up a significant proportion of the technical workforce. Theexact percentage, however, was hard to establish because data were scarce.Then, as now, most analysts lumped technical contractors together withother contingent workers, and estimates of the contingent labor force var-ied widely. The most liberal suggested that up to 30 percent of Americanswere contingently employed.22 The Bureau of Labor Statistics (BLS) of-fered a potentially more reliable and certainly more conservative estimate.According to their calculations, the percentage of Americans who workedcontingently between 1995 and 1999 stood at 13.3 percent.23

But even the BLS tabulated data in ways that made it difficult to esti-mate the number of technical contractors. Technical contractors werespread over several of the BLS’s categories of contingent workers. Peoplelike Kent Cox, who worked for themselves and who might or might notbe incorporated, were counted as “independent contractors.” YolandaTurner and Julian Stoke, who found work through staffing agencies,would have appeared as “employees of temporary service firms.” Stillother contractors, who were assigned to clients by consulting and out-sourcing firms, would have been labeled “workers provided by contractfirms.” To muddy the waters even further, incorporated contractors who

16 CHAPTER ONE

22 To arrive at such a high percentage, analysts count part-time work and all self-employed people as contingent workers, an assumption that seems difficult to justify (Dil-lon 1987; Belous 1989; Kalleberg et al. 1997). Most part-time workers, for example, arepermanently employed; they simply work less than forty hours a week.

23 See Polivka (1996a, 1996b) and Cohany (1996). Under its least restricted definition,the BLS defined the contingent workforce as the sum of (a) all wage and salary workerswho “do not expect their employment to last,” except for those who planned to leave theirjobs for personal reasons; (b) all “self employed (both the incorporated and the unincorpo-rated) and independent contractors who expect to be and had been in their present assign-ment for less than 1 year”; and (c) temporary help and contract workers who “expected towork for the customers to whom they were assigned for one year or less” (Cohany et al.1998, 43–44). The BLS repeated its assessment of the contingent workforce in the February1997 and February 1999 supplements to the Current Population Survey. The number ofworkers in the BLS’s broadest definition of contingent labor declined by 0.5 percent be-tween 1995 and 1997 and then stayed the same between 1997 and 1999. The percentage ofAmericans in each of the four alternative employment relations (independent contracting,on-call workers, temporary help agency workers, and workers provided by contract firms)largely remained constant except for a slight fall of 0.4 percent in independent contractors(Hipple 2001). Since three data points do not make a trend, it is difficult to determinewhether the decline represents random variation or an actual shift away from contingentlabor. The BLS’s data also cannot tell us whether contingent labor increased prior to 1995,even though most commentators contend that it did.

identified themselves as self-employed individuals escaped being catego-rized as contingent workers entirely. Nevertheless, the BLS data providea sense of technical contracting’s scope.

In 1997, roughly 20 percent of all independent contractors were profes-sional or technical workers, the two occupational categories into whichmost technical contractors fell.24 Professional and technical occupationsaccounted for 12 percent of all temporary help placements and 27 percentof the workers provided by contract firms. In fact, 23 percent of all contin-gent workers held professional and technical occupations in 1997. Draw-ing on the 1995 BLS data, Roberta Spalter-Roth and her colleagues esti-mated that 13.5 percent of all professionals in the United States eitherworked through staffing agencies, were independent contractors, or wereself-employed.25 Depending on whom one believes, in the late 1990s con-tractors may have accounted for 15 to 30 percent of the technical work-force in the Silicon Valley.26 Technical contractors were also common inother cities. In fact, they comprised the majority of some occupations—forinstance, multimedia designers in New York City.27

Although most commentators believed that contingent work and tech-nical contracting had grown exponentially during the 1980s and 1990s,lack of data also hampered estimates of growth. Before 1995 even theBureau of Labor Statistics did not collect information on the size of thecontingent workforce. In the absence of data on trends in contingent employment, most analysts used data on the growth of the temporaryservice industry as a proxy for the spread of contingent employment.These data show two indisputable trends. First, the temporary service in-dustry has grown spectacularly over the last two decades. Between 1986and 1996, employment in temporary services grew 10.3 percent, whiletotal employment grew only 1.7 percent.28 Second, the distribution ofcontingent work changed. Between 1991 and 1996, the percentage of thetemporary service industry’s payroll represented by office, clerical, andmedical work fell, while the industrial, technical, and professional seg-ments expanded.29 Clinton reported that between 1983 and 1995, thepercentage of programmers who found work through the business ser-vices industry grew from 24 percent to 40 percent.30 The percentage of

UNLIKELY REBELS 17

24 Cohany (1998).25 Spalter-Roth et al. (1997).26 Benner (1996).27 Batt et al. (2001).28 U.S. Department of Commerce (1997).29 Staffing Industry Report (1997).30 See Clinton (1997). The business services industry is broader than the temporary ser-

vices industry. It includes not only the temporary services industry but also firms that spe-cialize in providing outsourced services.

similarly employed computer engineers, computer scientists, and systemsanalysts increased from 18 percent to 29 percent.

Thus, three conclusions seem reasonable on the basis of available data.By the end of the century, a significant proportion of Americans werecontingently employed, the proportion had increased during the 1980sand 1990s, and technical professionals comprised a larger portion of thecontingent labor force than in the past. The trend seemed clear. LikeKent, Yolanda, and Julian, more and more technical professionals wereleaving the ranks of the permanently employed to work as itinerant ex-perts. The question was how to make sense of the trend. There were twoready-made alternatives.

Making Sense of Contingent Work

In the early 1990s, contingent work began to attract the attention of so-ciologists, labor economists, and journalists. Most commentators fo-cused on the reasons firms employed contingent workers and on thecosts and benefits of doing so.31 Analysts who took the workers’ perspec-tive were fewer in number and fell into two opposing camps: the “insti-tutionalists” and the advocates of “free agency.”

The Institutional Perspective

The institutionalists were economists or sociologists who interpretedcontingent work through a historical and social lens.32 They told a cau-tionary tale. From an institutional perspective, contingent work’s expan-sion threatened the security of the workforce and the American system ofsocial welfare, which was based on full-time employment. The institu-tionalists framed the threat from the perspective of dual labor market

18 CHAPTER ONE

31 We shall address why firms employ contractors in chapter 2. The research literature onfirms’ strategic reasons for employing contractors includes Abraham (1988), Abraham andTaylor (1996), Davis-Blake and Uzzi (1993), Harrison and Kelley (1993), Kalleberg andSchmidt (1996), Matusik and Hill (1998), Nollen and Axel (1998), Pfeffer and Baron(1988), and Rubin (1996).

32 “Institutional” and “institutionalist” are words with rich and thick meaning in the so-cial sciences. We use the term “institutionalist” to refer to analysts who view markets as“socially embedded” and who ask how institutions shape labor markets. By “institutions”they typically mean not only laws and established organizations but also cultural normsand practices. Economists who study how social structures affect markets have historicallybeen called “institutional economists.” Osterman (1988), Parker (1994), Barker and Chris-tiansen (1998), Cappelli et al. (1997), Carre and Joshi (1997), Kalleberg et al. (1997), andSmith (1998) are institutionalists who have written about contingent labor.

theory, the notion that industrial economies are composed of two sec-tors: “primary” and “secondary.”33

Primary labor markets provide stable employment, career ladders, jobsecurity, high wages, and attractive benefit plans. Secondary labor mar-kets are, by comparison, less stable and offer lower wages. Classic exam-ples of secondary labor markets include farm labor, food service, andhospitality work. Research has shown that participants in secondarymarkets are more likely to be members of minority groups and to workfor employers who provide meager benefits. Because scholars saw sec-ondary markets as peripheral, they had historically treated them as a so-cial problem to be addressed using existing institutions (for example,minimum wage laws), rather than as a fundamental threat to the systemof employment.

Institutionalists warned that the growth of contingent work representedthe spread of secondary labor into the economy’s core. Many feared thatthis development would undermine the well-being of American workersand their families.34 Others warned that contingent work’s spread would in-crease demand for government assistance in a downturn and facilitate theoppression of minorities. Indeed, demographic studies consistently showedthat most contingent workers made less money than full-time employees inthe same occupation, that they were less likely to have access to health in-surance and pension plans, and that in comparison to the full-time workers,women, African-Americans, and Hispanics comprised a larger percentageof the contingent work force.35 Finally, some critics charged that the shift tocontingent labor signaled an attempt to undermine unions.36 In short, insti-tutionalists saw the spread of contingent labor as an unraveling of the NewDeal, and they urged policy makers to bolster existing institutions or searchfor new ones that would enhance security.

Anthropologists and sociologists who had done ethnographies of contingent work generally confirmed the institutionalists’ fears.37 Their

UNLIKELY REBELS 19

33 For extended discussions of dual labor market theory, see Berger and Piore (1980),Piore and Sable (1984), Osterman (1984), and Baron and Bielby (1984).

34 For discussions of the negative social consequences of an expanding contingent work-force, see Dillon (1987), Osterman (1988), Martella (1991), Cohen and Haberfeld (1993),Hipple and Stewart (1996), Polivka (1996b), Christensen (1998), Barker (1998), Spalter-Roth and Hartmann (1998), Banigin (1998), and Houseman and Polivka (1999).

35 See Spalter-Roth et al. (1997), Kalleberg et al. (1997), and Kalleberg, Reskin, andHudson (2000) for demographic data on the contingent workforce.

36 Aronowitz and DeFazio (1994) and Rifkin (1995) level charges of union busting.37 An ethnography is a study of a group of people in which the data are collected by some

combination of participant observation and interviews. The ethnographer’s objective is usu-ally to describe and depict the “native’s point of view,” the perspective of the people studied.Parker (1994), McAllister (1998), Rogers (1995, 2000), Henson (1996), and Smith (1998)have written important ethnographies of contingent work in clerical and industrial settings.

studies depicted the difficulties that clerical and light industrial “temps”encountered in a world geared to full-time employment. Informants inthese studies reported being forced into temporary employment bydownsizing and other circumstances that made it difficult to find full-time jobs.38 Most complained of poor working conditions, low wages,and high work-related expenses. They spoke of disputes with clients andagencies over payment and hours, antagonism with permanent employ-ees, continual insecurity and uncertainty, and a sense of exclusion, es-trangement, and dissatisfaction with work.

The ethnographies reported few compensating advantages. Sometemps claimed to enjoy the flexibility of scheduling their own work andthe freedom to reject particularly unpleasant jobs. Others said they pre-ferred to receive their compensation as “fast cash” rather than wait for amonthly paycheck. Still others reported obtaining satisfaction fromknowing that their services were “really needed” by companies in crisis.But, in general, the ethnographic evidence indicated that at least amonglow-skilled workers, the disadvantages of “temping” outweighed its ad-vantages.39 It is significant that the institutionalists built their interpreta-tion of contingent work without examining highly skilled workers.40 Insharp contrast, commentators who wrote for the popular press largelybuilt their interpretations by ignoring the low end. Unsurprisingly, theypainted a considerably different picture of contracting.

The Free Agent Perspective

At the height of dot-com mania, “new economy” magazines like FastCompany and Wired began to lionize highly skilled contractors as “freeagents,” the heroes and heroines of postindustrialism. Most of freeagency’s advocates were (and continue to be) futurists, human resourceconsultants, and staffing industry experts who wrote for the general pub-lic.41 The “free agenteers” agreed with the institutionalists on one point:

20 CHAPTER ONE

38 Jurik (1998) is a notable exception in that only 20 percent of her self-employed homeworkers felt forced into their home businesses. Nevertheless, Jurik fixes on the perceptionsof those 20 percent in assessing how home workers feel about their work arrangements.

39 For exceptions, see Barker and Christensen (1998), Jurik (1998), Smith (2001), andRogers (2000).

40 Some institutionalists have suggested that the situation may be different for highlyskilled contractors, but their data do not allow much exploration. This is largely becauseeconomists and industrial relations researchers have relied on aggregate data and randomsamples that are heavily weighted toward the responses of traditional temporary employees.

41 Free agency’s spokespersons include Bridges (1994, 1995), Pink (1998, 2001), Beck(1992), Caulkin (1997), Darby (1997), Reinhold (2001), and McGovern and Russell(2001). Although Bridges was the first to market with the idea, Pink coined the name andgained notoriety. Pink runs a Web page called “Free Agent Nation” (http://www.freeagentnation.com).

employment security and its supporting institutions were unraveling andtheir demise represented a breach of faith. Corporations had reneged ontheir part of a culturally well-understood bargain. But, unlike the institu-tionalists, free agency’s spokespersons sang contracting’s praises whileadvocating a kind of libertarian, anticorporate rebellion.

Authors of this ilk portrayed corporate life as stifling and petty. Thecorporate world, they claimed, forced people to play “politics” and sub-ject themselves to the whims of incompetent managers for inadequatepay. Besides, they added, “jobs” and “careers” were outmoded inven-tions of the industrial revolution, designed for the benefit of employers.Daniel Pink, a former speechwriter for Secretary of Labor Robert Reichand the best-known advocate of free agency, wrote:

The old social contract didn’t have a clause for introspection. It wasmuch simpler than that. You gave loyalty. You got security. But nowthat the old contract has been repealed, people are examining bothits basic terms and its implicit conditions. Free agents quickly real-ized that in the traditional world, they were silently accepting an ar-chitecture of work customs and social mores that should have crum-bled long ago under the weight of its own absurdity. From infightingand office politics to bosses pitting employees against one another tocolleagues who don’t pull their weight, most workplaces are in dys-function. Most people do want to work; they don’t want to put upwith brain-dead distractions.42

Sometimes bordering on demagoguery, free agency’s antiorganiza-tional rhetoric reframed employment history. Firms had not simply bro-ken the contract; the contract was terrible in the first place because itlulled employees into dependency. Traditional employment and the ca-reers it offered were relics of corporate “paternalism” that encouragedemployees to play the role of children and firms, the role of parent.Simon Caulkin, a journalist writing for the San Francisco Examiner,explained:

At the bottom, the trade of loyalty for security was unsustainableand exploitative. By tying their career and skills to one employer,employees sacrificed mobility and market value in return for apromise that couldn’t be kept. By contrast, in the emerging newdeal, the very mobility of career-independent employees provides apowerful incentive to companies to keep their promises—the mostimportant of which . . . is to provide interesting, motivating work.Otherwise, those valuable staff members will leave. So the death ofthe security-loyalty contract has made room for a more durable and

UNLIKELY REBELS 21

42 Pink (1998, 132).

satisfying employment relationship—this time based on indepen-dence rather than the employer paternalism of the past.43

Faced with the demise of traditional employment and the potential forfirms to renege on their part of the bargain, advocates of free agency rec-ommended that people turn the tables on companies by refusing to giveloyalty and by embracing the shift to a skills-for-hire economy. The trick,they proposed, was to use the demise of traditional employment as anopportunity to set oneself free. They encouraged people to view them-selves as free agents, to develop and market their own skills to the high-est bidder, and to view themselves as a business, even when in a full-timejob. Free agenteers, in short, promoted a postindustrial vision of eco-nomic individualism in which entrepreneurial workers would regain in-dependence and recapture a portion of their surplus value.

To support their claims, free agenteers offered readers stories of con-tractors who vacationed when and where they chose, who telecommutedto work from exotic places, and who successfully integrated the demandsof work and family. Although the heroes and heroines of these tales wereusually professional and technical workers, proponents contended thatall people could benefit by adopting a similar attitude to work. The cen-tral icon of the movement was the hero of the movie Jerry Maguire, whorejected a world of greed and unethical behavior for independence,wealth, and self-respect. Caught up in the revolutionary spirit, Pink andWarsaw even penned what they called a “Free Agent Declaration of In-dependence.”44

The vehicle of the free agent’s independence was the market. In JobShift, William Bridges promised, “In a market . . . people don’t havebosses. . . . There are no orders, no translation of signals from on high,no one sorting out the work into parcels. In a market one has customers,and the relationship between a supplier and a customer is fundamentallynon-organizational. . . . One’s boss is really a major customer rather thanan authority in the old sense.”45 Of course, the irony that Bridges andothers had trouble seeing was that freeing oneself from the chains of tra-ditional employment meant catering to the same employers againstwhom one was rebelling.

Despite such inconsistencies, the advocates of free agency sketched an

22 CHAPTER ONE

43 Caulkin (1997, 2).44 The “truths” that Pink and Warsaw (1997) claim to be “self-evident” are: “Who we

are and what we do should not stand on opposite sides of a psychological divide. . . . Noth-ing is permanent. Security is an illusion. . . . The power to choose is the power to say no.. . . Fear has no place in Free Agent Nation. . . . The fun in work is the reason for work. . . .We’re on our own, but we’re not alone.”

45 Bridges (1994, 64–65).

optimistic picture of contingent labor based on a rhetoric of free marketsthat challenged the institutionalists point by point. Free agenteers arguedthat contingency was a choice rather than a constraint; that it repre-sented liberation rather than isolation; that it minimized uncertaintyabout employment while enhancing flexibility and personal control; thatcontractors made more money than permanent employees because theywere paid at rates that reflected the real value of their skills; and that re-lying on one’s skills led to self-actualization rather than estrangement.

Who to Believe?

As sociologists with a background in industrial relations and a taste forleft-of-center politics, our initial response was to side with the institu-tionalists and to treat the rhetoric of market freedom with a healthy doseof disbelief, if not suspicion.46 But as ethnographers, we believed thatdata should be allowed to speak for itself. No a priori formulations, nomatter how elegant or consistent with their proponents’ worldview,should be exempt from the requirement that they rest on solid empiricalevidence. For us, evidence meant an empathetic and rich description ofthe perspectives and practices of the people about whose lives social sci-entists made claims.47 From this perspective we found both sides of thedebate wanting.

How could anyone tell whether the institutionalists or the free agen-teers offered a more viable image of contracting when neither had sys-tematically examined the lives of technical contractors? With the excep-tion of the ethnographers of temporary work and Daniel Pink’s book,Free Agent Nation, analysts on both sides of the debate had given novoice to contingent workers themselves. Most advocates of free agencybased their claims, at best, on well-chosen anecdotes. The institutional-ists were much better empiricists, but their data were heavily weightedtoward the experiences of low-skilled temps. This bias was especially

UNLIKELY REBELS 23

46 In fact, we had already written a critical historical analysis of how American manage-rial rhetorics have repeatedly vacillated since the 1870s between rhetorics of rational andnormative control (Barley and Kunda 1992).

47 The ethnographer’s creed is that all interpretations of a social system must rest on “thenative point of view,” or what anthropologists have called an “emic” perspective. The dis-tinction between emic and etic is the anthropologist’s way of distinguishing concepts usedby the people being studied and those used by the people doing the studying. The termscomes from linguistic anthropology. “Etic” is a term derived from the word “phonetic.” Itcontrasts with “emic,” which comes from “phonemic.” “Phonetics” is the study of thesounds that people can physically produce. “Phonemics” is the study of the sounds thatpeople distinguish as meaningful. Others have used the terms first–order and second–orderor experience near and experience far to make the same distinction.

troubling because students of work have long known that the worlds oflow- and high-skilled occupations substantially differ. Permanent profes-sional, technical, and managerial jobs are usually more secure, more re-munerative, more varied, and more involving than clerical and industrialjobs. It stands to reason that low- and high-skilled contingent workshould vary in similar ways. By ignoring highly skilled contingent work,institutionalists conflated the effects of contingent employment with cor-relates of low-skilled work. Conversely, advocates of free agency peddledimages of contracting built solely on the experiences of an elite. In theabsence of adequate data, the institutionalists’ and the free agenteers’claims about the everyday realities of contracting and the theoretical ed-ifices they erected on these claims, rang hollow.

We had substantive concerns as well. These rested on our reading ofthe existing literature, on our experience as researchers of technicalwork, and on our anecdotal familiarity with technical contracting. Meet-ing contractors like Kent, Yolanda, and Julian reinforced our concerns.Three issues, in particular, troubled us.

First the discourse on contracting seemed too rigid, too black andwhite, to be believable. For institutionalists, contingent work was a clearand unambiguous social problem. Temps and contractors were victims ofsystemic changes promulgated by exploitative employers acting entirelyin their self-interest without regard for the common good in the face of agovernment that was unable or unwilling to protect them. For the advo-cates of free agency, contingent employment represented no less than apath for escaping a decaying system that had subjugated the many to thewhims of the few. If ethnographers have learned anything about sociallife it is that reality rarely comes so neatly packaged.

Which camp’s box, for instance, could comfortably hold Kent Cox,Yolanda Turner, and Julian Stoke? The institutionalists could easilyargue that Julian’s experience supported their contentions. He became acontractor after being laid off. He had clearly lived on the edge of finan-cial solvency. He had difficulty making ends meet, he had trouble keep-ing a job, his wife was anxious about their future, and he vacillated be-tween contracting and permanent employment. Even protecting hisfamily’s health had been a serious problem. But, to the institutionalist’sconsternation, Julian didn’t seem to mind his troubles! Worse yet, heconsidered himself to be a success, at least in his own terms, and heplanned to continue contracting.

Kent Cox would give the institutionalists an even bigger headache.Other than missing the opportunity to cash in on the fantasy of becom-ing rich through stock options, Kent seemed in no way disadvantaged. Infact, by all indications, and certainly in his own view, contracting hadbeen good to him. But it would be Yolanda Turner who most violated the

24 CHAPTER ONE

institutionalist’s worldview. Here was an African-American woman who,despite her efforts, had been unable to flourish until she took her career,her skills, and her wages into her own hands as a contractor.

On the face of it, Yolanda’s and Kent’s tales should have brought greatcheer to free agency’s camp. Either could have easily been a sidebar inone of the slick “new economy” manifestos on free agency. Yet, on closerscrutiny, their stories posed problems. At least in Yolanda’s case, freeagency’s promise of freedom, flexibility, and leisure seemed elusive. Infact, Yolanda worked long and hard, at least as much and maybe morethan she had as a full-timer. Her employability depended uncomfortablyon staffing agencies, and her potential earnings were hijacked by theirmarkups.

But Julian would have been free agency’s real embarrassment. EvenWired would have had to struggle to find glamour in an overweight,balding middle-aged man with credit card problems and a penchant forbeing fired. Moreover, Julian lived hand to mouth and suffered seriouslyfrom the lack of health care benefits. From an ethnographer’s per-spective, a satisfying analysis of contracting has to account for all threestories.

Second, except for the ethnographers of clerical work, most analystson both sides of the debate spoke as if contingent labor markets weredyadic: composed simply of employing organizations and contingentworkers. Although this is sometimes the case when a contractor is incor-porated or works as an independent, in most instances, contingent labormarkets are triadic. In addition to buyers and sellers there are brokers,the staffing agencies that mediate between the two to arrange and closedeals. Contractors, like Yolanda and Julian, regularly rely on staffingagencies to find work. An accurate analysis of contracting must, there-fore, take the agency’s role into account regardless of whether the goal isto portray the structure of the market or describe contractor’s experiencemore fully. Leaving out such an important group of actors guaranteesmisrepresentation.

Finally, the debate on contingent work was framed as if it were an ar-gument between proponents of organizations and advocates of freemarkets. Institutionalists set their vision of employment against thebackdrop of an organizational society, its laws, and its regulations.When institutionalists recommended reform, they inevitably turned tolegal and bureaucratic remedies for the social costs of unfettered mar-kets. They believed that full permanent employment was the foundationof an equitable system. Conversely, advocates of free agency preferredthe market as a model for organizing the world of employment. Forthem releasing the forces of a free market would remedy the ills of anoverly bureaucratic society. While organizations and markets are clearly

UNLIKELY REBELS 25

important for framing an understanding of and prescriptions for theworld of employment, they do not exhaust the possibilities. Occupa-tions, which both institutionalists and advocates of free agency ignore,are just as important.

Sociologists of work have repeatedly shown that lawyers, scientists,police, carpenters, machinists, accountants, and members of other skilledoccupations construct their identities and organize their practice not onlyaround their employer or the market for their services, but also aroundtheir occupational affiliations.48 Ask a doctor what she does for a livingand she’ll first tell you her specialty, not that she works for Kaiser Per-manente. This was precisely what Kent, Yolanda, and Julian appeared tobe saying when they highlighted the importance of their networks andtheir involvement in professional communities, regardless of whetherthey found those communities at a users’ group, through a professionalorganization, or in a chat room.

Professionals’ closest contacts are usually other practitioners of theircraft. Professional associations, occupational unions, and communi-ties of practice provide experts with critical support, knowledge, andinformation. In fact, the more esoterically skilled the practitioner, themore an occupational community is likely to be important. Sincetechnical contractors were highly skilled, it seemed to us that their occupational affiliations had to be at least as relevant as markets and organizations for understanding contracting and interpreting itssignificance.

The Study

It was to address our methodological and substantive concerns that weentered the world of contracting in the fall of 1997 and remained thereuntil the fall of 1999. But getting inside was a little more complicatedthan is usual for ethnography. Doing fieldwork in an organization, a sub-culture, or a tribe requires entering a single, relatively well-delineated so-cial setting, which, as most ethnographers discover, is difficult enough.Getting inside the market for technical contractors meant understandingthe story from three perspectives: the contractor’s, the client’s, and thestaffing agency’s. It also meant studying a sprawling, emergent, looselycoupled, and at least partly virtual social system.

26 CHAPTER ONE

48 We offer an extensive argument to support the importance of an occupational per-spective in organizational analysis in Barley and Kunda (2001).

Staffing Agencies

We turned first to staffing agencies, in part, because their role in contin-gent labor markets is poorly documented and, in part, because as bro-kers they continually peer into the other two worlds. By making friendswith agents we hoped to make the contacts necessary for branching outinto the other two worlds.

In November 1997 we began six months of participant observation inProgressive Staffing, a medium-sized staffing agency that specialized inplacing technical contractors ranging from workstation technicians andtechnical writers to programmers and hardware engineers. Headquar-tered in Santa Clara County, Progressive Staffing maintained a numberof field offices throughout the Silicon Valley as well as offices in eightother American and European cities. Founded in 1993 by Anne Parish,an energetic entrepreneur whose vision included a commitment to run-ning a humanistic corporation that treated contractors well, ProgressiveStaffing had gone public shortly before our arrival. During our sixmonths there, we observed close-up for several days each week the tribu-lations and jubilations of matching contractors to jobs. We observed in-terviews, eavesdropped on telephone conversations, accompanied man-agers to client sites, read materials, studied computer screens, attendedmeetings, took part in training programs, and struggled to remain soberat corporate celebrations.

After completing our work at Progressive Staffing, we began similarthree-month stints at two other staffing agencies, chosen because theirbusiness models differed significantly from Progressive Staffing’s. Infor-mation Technology Specialists (ITS) was much smaller than ProgressiveStaffing and specialized in placing independent IT contractors who wereincorporated as businesses. ITS had one office, located in San MateoCounty, California. Marc Sunberg, a former systems analyst, had foundedthe firm in 1980. The firm had evolved through a number of namechanges as well as changes in strategic direction. The contractors whoworked through Progressive Staffing became Progressive Staffing’s em-ployees for the duration of their contract, and that was the modal pat-tern for most other agencies. But ITS was different. It preferred to have acorporation-to-corporation relationship with its contractors.

Systems Professionals, the third agency, was also smaller than Progres-sive Staffing and even more specialized than ITS. Systems Professionalsconcentrated exclusively on providing UNIX and Windows systems ad-ministrators to firms in the Silicon Valley. It differed from ProgressiveStaffing and ITS in that it operated as an outsourcer. All but a handful of

UNLIKELY REBELS 27

Systems Professionals’ contractors were full-time employees who re-mained employees even between contracts. Most were young and freshout of college where they had been responsible for running computerclusters. Because they were outsourced employees, Systems Professionals’contractors could work for clients for long periods of time and they oftendid, even though they could request and receive a new assignment when-ever they desired. Technical personnel tended to remain with SystemsProfessionals for two to three years before leaving for another permanentjob. Just before we arrived, Systems Professionals had opened a branchoffice in New York City.

Contractors

During 1998 and 1999 we interviewed seventy-one contractors abouttheir careers and experiences. Kent, Yolanda, and Julian were amongthem. Because there were no truly representative lists of technical con-tractors from which to draw a random sample, we identified our inform-ants in three ways.49 Half (thirty-six) were selected from a list of nearlyfive hundred contractors who indicated their willingness to take part inour study when registering for a seminar on contracting business issuesthat was broadcast live over the World Wide Web in December 1997. Theseminar was sponsored by Progressive Staffing and was widely advertisedthroughout the nation in contracting circles via a number of channels in-cluding magazines written for technical contractors. To ensure that ourconclusions were not regionally biased, we actively sought to talk to con-tractors from outside the Silicon Valley who had registered for the semi-nar. We also made a particular effort to interview women. These originalinformants referred us to five other contractors whom they thought wewould be interested in interviewing for one reason or another—usuallybecause an informant thought a friend’s experience of contracting wasdifferent from their own. We met another fourteen contractors while wewere doing participant observation in the staffing agencies. The remain-ing sixteen contractors were members of the project teams we encoun-tered when studying client organizations. Although our informants werenot representative in a statistical sense, they spanned numerous regions,occupations, and age cohorts as well as both genders. For this reason, weare reasonably confident that our data identify key themes, issues, anddilemmas that are of widespread concern to technical contractors.

28 CHAPTER ONE

49 As we saw it, our alternatives were to convince a staffing firm to make available thenames of the people in their databases, to sample from one of several résumé data banksfound on the Internet, or to seek subscription records from magazines targeted at contrac-tors. All of these approaches are biased in different ways and seemed no better or worsethan the approach we took.

When possible we met our informants in restaurants, cafés, offices,and even their own homes.50 When it was impossible to do face-to-faceinterviews, usually because a contractor lived in another state or becausehe or she did not wish to meet, we conducted telephone interviews. Ourinterviews covered the contractors’ employment history, the circum-stances that led them to become contractors, their rates, their experiencewith staffing firms, how they found jobs and remained up-to-date, theirexperiences with clients, the impact of contracting on the rest of theirlives, and their perceptions of the costs and benefits of being a contractor.With the exception of two contractors who did not wish to be recorded,we taped and transcribed all interviews. Most of the interviews lastedfrom one and a half to three hours.

Client Firms

The final phase of our study took us to the clients’ world: firms in the Sil-icon Valley that routinely employed technical contractors. Three of thefirms were among the largest and best known in the computer industry.In each of these firms we identified two project teams staffed by a combi-nation of contractors and full-time employees. All of the teams were in-volved in software development or information technology projects. Weinterviewed the permanent employees and contractors on each team aswell as the managers who ran the projects and who had hired the con-tractors. In one firm, we interviewed all the way up the hierarchy to thechief information officer (CIO).

Our interviews with permanent employees concentrated on theirwork, their experience with contractors, and their view of the role thatcontractors played in their firm. When talking to managers we focusedon the firm’s policies and procedures for hiring and terminating contrac-tors, the costs and benefits of using contractors, the manager’s experi-ences working with agencies, and the challenges of managing a mixedworkforce.

We extended our study of clients by interviewing managers at sevenother firms. Two were large firms in the telecommunications industry.The rest were entrepreneurial start-ups that employed no more than onehundred people, including contractors. Because these organizations weresmaller, the managers tended to have broader responsibilities than the

UNLIKELY REBELS 29

50 Anthropologists use the term “informants” to describe the people whose perspectivesthey study. They prefer the term because it underscores the idea that the ethnographer’s jobis to learn to see through the insider’s eyes. “Informant” contrasts with the idea of a “sub-ject” or a “respondent,” terms that are favored by survey researchers and experimentalists,respectively. We use the term “informant” throughout the book to remind the reader thatthe people we studied taught us everything we know about contracting.

managers that we interviewed in the project team studies. All but threeinterviews were conducted face-to-face at the client’s site where we couldobserve working conditions.

Collecting and analyzing these data consumed the better part of fouryears. By the time we were done, it was clear that the world of technicalcontracting was more complex than the institutionalists or the advocatesof free agency suggested. Both perspectives capture a piece of reality, butneither adequately portrays contracting as those who do it experience it.This limits their ability to offer more comprehensive images of how con-tracting fits into the larger world of employment. Our goal is to providea window into the day-to-day realities of contracting as a basis for devel-oping a wider interpretation of contracting’s relevance in a knowledgeeconomy.

History, however, does not wait for ethnographers to complete theirwork. About a year after we left the field, as we were pondering the sig-nificance our data, the economic conditions of high-tech employmentchanged dramatically. The Internet boom and the tight labor market itspawned collapsed, ushering in a recession whose effects are still beingfelt at the time of our writing. Along with the rest of the economy, theworld of contracting shifted as well. In short, our tale portrays contract-ing at a peak of demand. Whether it speaks only to a specific historicalperiod, and the extent to which it continues to be relevant, ultimately re-mains to be seen. This, of course, is the fate of all social science. But webelieve that the dynamics we discuss over the course of this book lie be-yond the short cycles of the economy: they define the essence of con-tracting as an employment relationship. In the book’s epilogue, we dis-cuss what has occurred since the recession began and how it affected ourinformants’ world. The data show that like full-time employees, contrac-tors have suffered, but they have not disappeared. On the contrary, theirpresence remains a significant feature of contemporary labor marketseven in times of recession. Thus, with appropriate modifications forchanging levels of demand, our analysis of the dynamics in the labormarket of the 1990s still holds.

Organization of the Book

What follows is the story of contracting told from the participants’ per-spectives, especially the perspectives of contractors themselves. In tellingthe contractor’s story, we shall also have much to say about the agenciesthat place contractors, the clients for whom contractors work, the net-works in which contractors move, and the supports on which contrac-

30 CHAPTER ONE

tors rely, ranging from magazines to the Web sites and users’ groups thatlubricate the systems’ operation.

Like all ethnographies, ours rests on the details of what we saw andheard, the patterns we found in our data, and the sense we have made ofthose patterns. To evaluate our claims, readers also need to have famil-iarity with the data. Accordingly, we tell our tale from the bottom up.Our telling, therefore, differs from the storyline of most academic andbusiness books, which hammer home their main points in the first chap-ter, often with rhetorical certainty and the desire to relieve the reader ofreading any further. The remaining chapters then unpack the conclusionpoint by point, providing further justification and sometimes supportingevidence. Rather than starting with conclusions, we shall take you on aguided journey through the world of contracting to give you an appreci-ation for how contracting appears to those on the ground and in thetrenches. By presenting a series of grounded portraits, we help you to as-semble your own collage as we go along. By the time we are ready todraw together our interpretations of the social world of contracting, youshould be in a position to evaluate the adequacy of both our analysis andour conclusions. For readers eager for the security of maps before enter-ing the terrain, we offer the following overview.

The book unfolds in four parts. Part 1 sets the stage. In it we definekey terms, introduce the main characters, and examine their motivations.In chapter 2 we explore why firms hire contractors and how they orga-nize themselves to do so. In chapter 3 we explain why technical profes-sionals become contractors and discuss the various ways contractorsorganize their practice. In chapter 4 we present a typology of staffingagencies and explain how they operate.

Together, parts 2 and 3 depict a turn in the “contracting cycle,” whichis composed of a period of searching for a job followed by a limited en-gagement with a client. Repetitions of the contracting cycle provide therhythm against which contractors’ lives unfold. Part 2 immerses thereader in the market. In chapter 5 we tag along as contractors, clients,and agencies try to find deals. We look over their shoulders as they scourthe market for potential deals and disseminate information about theirown interests. In chapter 6, we see how deals, once found, are consum-mated. We examine how clients evaluate candidates, how terms of em-ployment are negotiated, and how contracts are signed. By the end ofpart 2, the contractors have acquired a contract.

In part 3 we follow the contractors onto the job. Chapter 7 shows howmanagers and staffing firms attempt to create a system for managing con-tractors that is consistent with corporate policies designed to minimizelegal risks. Chapter 8 points to the inherent tensions between corporate

UNLIKELY REBELS 31

policies and the realities of technical work. Chapter 9 shows how con-tractors navigate between the respect and the resentment that their roleengenders among managers and permanent employees to carve out a co-herent identity from their experiences.

Part 4 turns to how contractors adapt their lives to the requirements of the contracting cycle. It focuses on the way they develop, organize,maintain, and use three resources crucial for successfully riding the cycle:time, skills, and social relations. Chapter 10 examines the temporalrhythms of contracting, how contractors treat time as an asset, and howthey allocate their time to the various requirements of a contracting life.Chapter 11 explores how contractors think about skills and how theykeep themselves up-to-date and marketable. Chapter 12 focuses on howcontractors build their social networks and how they use them to repeat-edly find work.

The book ends with chapter 13. Here we summarize our sojournthrough the world of contracting and draw together what we havelearned. We then use our findings, in the manner of ethnographers, tobuild a theoretical interpretation from the bottom up. Not straying toofar from the data, we first develop the claim that contractors pursue aform of professional practice that we call “itinerant professionalism.”We discuss the attributes of this form of practice by contrasting it toother forms of professional practice, and explore its significance for con-tractors, clients, and agencies. Then, in a somewhat riskier fashion, weinvite our readers to join us in pushing our conclusions even further byspeculating on the significance of itinerant professionalism for a knowl-edge economy. We explicate and critique the institutionalists’ and freeagency’s views of the economy in light of our data, highlighting the costof neglecting occupational forms of organizing. This critique leads us toargue that the social organization of work in a knowledge economy isusefully viewed as a matrix, which incorporates not only markets and or-ganizations, but occupations as well. We conclude by speculating brieflyon the relevance of this perspective for policy makers and practitioners.

32 CHAPTER ONE